Marker
Marker is a tool that converts PDFs, images, and office documents such as Word, PowerPoint and Excel into clean markdown, JSON, HTML or text chunks. It is aimed at programmers and IT people who need document text for other software.
How hard it is to start using it. Easy: install it like a normal app, or use it in your browser. Advanced: you run it on your own server, use the command line, or it is a tool for developers.
Features
Reads scanned pages
SourceYes
Keeps tables intact
SourceYes
Reads many file types
SourceYes
No commands neededHas a visual app, but install and setup need terminal commands
SourcePartly
Runs AI on your own computerLocal server possible, but needs Docker or llama.cpp setup
SourcePartly
Free for business useCode is free commercially; model weights limited to startups under $5M
SourcePartly
Quick to run yourselfPython install plus separate PyTorch and inference backend setup
SourcePartly
Listed in
Converts documents to text and chunks, but is a command-line tool, not a search service.
How to help
AI contributionsNot stated
The project has not said how it feels about AI-written contributions. Maintainers can set this by claiming the listing.